1 comment

[ 3.3 ms ] story [ 13.8 ms ] thread
Amazing overview of the various compression schemes used in columnar stores.

I use Parquet with ZSTD with DuckDB and it absolutely flies.

Have not reached for Spark in a long while.